Conversation starter: How heavily do you rely on computational tools?

[human909]

Conversation starter: How heavily do you rely on software computational tools?

---

In a effort to start an interesting conversation.... How heavily do you lean on computational tools. To help aid the discussion I'll try to break up the tools into categories in order of computationsal complexity: (I'm attempting to define groupings here, if things clearly don't fit feel free to let me know and I'll adjust...)

1. 3D FEA either in 3D bodies or 3D shells (including BUCKLING ANALYSIS)
2. 3D FEA either in 3D bodies or 3D shells
3. 3D analysis with 1D elements (eg; typical structural analysis programs where elements are defined as lines (columns/beams) in 3D space)
   
   1. Member design via software
   2. Member design via hand calcs or similar
4. EXCEL or similar calculative spreadsheets/software.
5. HAND calcs (calculator allowed)

---

 

[XR250]

I do mostly residential work and I put an entire project in one workbook. I haven't published hand calcs in more than 15 years.

I'm curious what your workflow would look like if you did not have to submit calcs.

 

[Lomarandil]

For general construction engineering:
3.2: 30% (RISA with MathCAD design calcs)
4: 65% (MathCAD base file with some excel as standard spreadsheets or when post-processing data)
5: 5% verifications of the above

For specialized bridge projects:
2: 15% (MIDAS, mostly to get staged construction and effects of skew/curvature, not for the 3D elements)
3.2: 60% (RISA and MIDAS)
4: 20%
5: 5%

 

[dauwerda]

1. 5% - IdeaStatica for some funky connections
2. 0% - (One could argue most of my IdeaStatica use is this)
3. 50% - RISA and PLS-Pole
4. 45% - Excel
5. 0% - I use my calculator every day to check stuff or when doing quick computations when modeling, but not in the sense of doing hand calcs on paper.

 

[ChorasDen]

My default is to use typical widely available design software, they are typically FEA due to the the software authors choices, but for the work I do, they do not strictly need to be FEA.

For non-typical situations where I will do the calc 5 or more times over the next few years, I will prep an entire example problem via a handwritten paper solution, with commentary explaining any derivations that are non-obvious, and write up a solution in a compile-able language (C++,C#, python if I hate myself). If I am feeling really energetic, I'll also include a methodology document, but that tends to take more time to write than the actual solution, but it also helps you to get back in the mindset if you've let the tool alone for a year+ and you need to remember how anything works.

 

[ChorasDen]

For various reasons, I have, and likely always will, hate using Excel. So I default to using a compiled language rather than trying to decipher what various cell references are trying to indicate in any one of the archaic spreadsheets that are out there. I also do not have a personal excel license, so I cannot typically continue work on a home machine if desired.

 

[271828]

...in a compile-able language (C++,C#, python if I hate myself)....

I'm curious what compiler/editor combination you use for C++.

 

[rb1957]

I'd've said aero is 50% 1 and 50% 4 ... in my experience there is a lot of hand cranking after the FEA has had it's go.

And then of course 50% testing ... all primary structures are tested at some point.

 

[ChorasDen]

I'm curious what compiler/editor combination you use for C++.

I don't write in C++ as much as I used to back in college, but I still do it from time to time if I am feeling there is a particularly good reason to do so. Working in OpenGL was the last significant work I did in C++, is OpenGL even still around anymore? Seems I could be dating myself. To answer your question, I honestly just use visual studio community. It works well for what I need it for, I like the git integrations, and I haven't spent much time branching out to other IDE's to see if there are any features that it's missing.

 

[lucky-guesser]

Pretty 50/50 between 4 (excel) and 5 (hand). That said, in my mind when you are using a company excel sheet or something written by a coworker, that is a separate category of "I hope this excel sheet formula was written correctly and gives good results". However, when I am using an excel sheet that I have written myself, in my mind I basically see it as a hand calc, it is just in excel format as to make life easier and more consistent for repeat calculations.

 

[271828]

... is OpenGL even still around anymore? ...

It is. https://www.opengl.org/

Looks like it's intended for high-end graphics applications. Not sure what it used to be.

 

[ChorasDen]

When I was working in it about 5 years ago or so, it was because OpenGL was open source and licensing was free (maybe only for non-commercial? I cannot recall), rather than something like DirectX, or the alternate Vulcan, which I was unfamiliar with. I was doing some work visualizing large scale spatial modeling using LIDAR datasets into an interactive 3d model in userspace, and I certainly did not want to build my own rendering pipeline.

 

[IDS]

I think there is a huge overlap between the categories. In particular Excel might be used for anything from category 1 through to 5. There is also still a large place for 2D analysis, especially for soil-structure interaction problems, and non-linear materials analysis should also be on the list.

My work almost always includes some FEA, mostly 2D with non-linear materials and modelling of the construction sequence. Non-linear geometric effects are included if the deflections are significant. Excel is nearly always used for generation of FEA input and analysis of the results, member design and generation of design output summary information (tabular and graphic).

"Hand calcs" if required (for instance for an independent check of FEA results or spreadsheet output) are always done in Excel. I have a calculator somewhere, but I don't remember the last time I used it, and I don't even know where it is now.

The spreadsheets use VBA extensively, but these days that is progressively being replaced with Python, mainly because of access to libraries like Scipy, and because having the code separate from the spreadsheets makes control and security a lot easier.

 

[NorthCivil]

I'm curious what your workflow would look like if you did not have to submit calcs.

ive practiced in areas where you need to submit calcs, and where you dont.



Where calcs need be submitted:

• takes more time (to format nicely and present calcs)
• there is pressure to calculate things that otherwise you could just judge as OK by inspection.
• there are peer reviewers snooping around having a pick at your work from time to time, killing profitability
• often times the peer review ends up being a de-facto training session (their benefit, your cost)
• generally there is less "working in silo" effect - work is being reviewed a lot more.
• cost of all this is higher.

Where no calcs need be submitted:

• engineering judgement skills become much sharper
• cost is lesser
• more "engineer in silo" effect
• more time is spent looking at the drawings vs packaging a nice tidy report for someone else to read.
• adequate documentation of work can be a chore, and it can get sloppy easily.

I prefer not having to submit calcs, myself.